Very exciting!
Looks like I was able to calculate my polygenic score for Educational Achievement.
This is still on the work bench stage, though my checks to date have checked out.
This should really be a 3 line code type problem, though I went somewhat off road to arrive at an answer.
Setup was I had the 5 split files. For each of the files, I went through all 1 million genotypes to find out if the genetic location was present. For the variants for EA there were ~500 (of 1,200) that were absent in all 5 files. This means that these variants were reference homozygous for me. I then needed to find out if these reference alleles were Allele 1 or Allele 2 in the polygenic score. Many of the reference were Allele 2 which meant they counted for 0. The reference genotypes that were Allele 1 were then worth twice the beta.
For the non-reference homozygous I just multipled the beta by the number of alternative alleles. (Will need to recalculate that one. Those genotypes that have 1 alternative and 1 reference clearly will have a beta equal to the beta value. For the homozygous alternative I will need to check whether alternative is actually Allele 1 in the polygenic score (if it is then I will need to double it, if not then I will just record it as 0).
The polygenic score appears to be calculatable, though the size of the files involved have made it somewhat more laborious than would be optimal. Nevertheless, these polygenic scores offer a very powerful method of exploring personal traits.
